A mmW image-based algorithm on wire recognition for DVE applications

• Main Authors: Goshi, Darren S, Sun, Ming-Ting, Kirk, John
• Format: Conference Proceeding
• Language: English
• Published: SPIE 05.05.2017
• ISBN: 9781510608955; 1510608958
• ISSN: 0277-786X
• DOI: 10.1117/12.2262250

In this work, we present the framework surrounding the development of a mmW radar image-based algorithm for wire recognition and classification for rotorcraft operation in degraded visual environments. While a mmW sensor image lacks the optical resolution and perspective of an IR or LIDAR sensor, it currently presents the only true see-through mitigation under the heaviest of degraded vision conditions. Additionally, the mmW sensor produces a high-resolution, radar map that has proven to be exceedingly interpretable, especially to a familiar operator. Seizing on these clear advantages, the mmW radar image-based algorithm is trained and evaluated against independent mmW imagery data collected from a live flight test in a relevant environment. The foundation of our approach is based on image processing and machine learning techniques utilizing radar-based signal properties and sensor and platform information for added robustness. We discuss some of the requirements and practical challenges of a standalone algorithm development, and lastly, present some preliminary examples using existing development tools and discuss the path for continued advancement and evaluation.